Do EU-MED agreements improve democracy and the quality

Do EU-MED agreements improve democracy and the quality of institutions in EU partner countries? Cindy Duc Emmanuelle Lavallée EURIs. CO Université Paris Dauphine 1

Purposes ¢ ¢ 2 To assess the effects of the EU-MED agreements on the respect of democratic principles and the promotion of “good governance” To investigate the timing of these effects

Motivations ¢ ¢ A recent trend in North South preferential agreements: non-trade conditionality; Particularly true for the EU: l 3 since the beginning of the 90’s the EU has included in all its trade agreements provisions connected with the respect of democratic principles and the promotion of “good governance”;

Motivations (continued) ¢ EU-MED agreements: l l l ¢ Barcelona declaration; MEDA program EU-MED agreements Particularity of EU-MED agreements: l Mainly autocratic countries èHas the “political conditionality” imposed by the EU been strong enough to allow people to enjoy more liberties? l No « obligation to achieve a result » èHas EU-MED agreements contributed to improving the quality of the institutions of Mediterranean countries? 4

Governance measures ¢ Two indicators of democracy: l l ¢ Polity IV Freedom House: political rights and civil liberties Governance indicators: IRCG l l l Quality of bureaucracy Corruption Law and order èAn overall governance indicator 5

The overall impact of EU-MED agreements • Yit is governance in country i at time t, • ai is a time-invariant effect unique to individual i • Tt is time effect common to all countries at time t, • Xit is a set of other control variables, • Euromedit is a binary Euro-Med membership indicator, it takes the value of 1 from the year the agreement was signed and 0 otherwise • eit is an unobserved error term 6

The timing in the effects • yit, ai, Tt, xit, eit are defined in the same way as in equation (1) ; • Euromed(-2)it is a dummy variable that takes the value of one the two years before the agreement was signed, and zero otherwise; • Euromed 13 it equals to one the year when the agreement was signed and the two following years; • Euromed>=4 it equals to 1 from the 4 th year after the agreement was signed. 7

Two estimations techniques ¢ Within comparisons: only use the time variation in the data. It consists in estimating equation (1) and (2) on a sample made only with countries that have signed an agreement. ¢ Difference-in-differences estimations (DID): consist in identifying a specific intervention or treatment. One then compare the difference in a dependent variable before and after the intervention for groups affected by the treatment to the same difference for unaffected groups. 8

Effects on democracy using Within Comparisons Polity 020 FH 1 2 3 4 5 6 7 8 -0, 24 0, 37 _ _ -0, 32 0, 5 3* * _ _ Euromed(-2) _ _ 0, 16 0, 52 _ _ -0, 32* -0, 53* Euromed 13 _ _ 0, 03 0, 67 _ _ -0, 42 0, 72 *** Euromed>= 4 _ _ 0, 61 1, 92* _ _ -0, 38 -1, 01** Ln(GDP/capita) _ 7, 75 *** _ 9, 21 *** _ 1, 91* _ 1, 83** % Fuel exports _ -0, 02 _ -0, 01* _ 0, 02 *** Euromed 9

Effects on democracy using DID Polity 020 1 2 3 4 5 6 -1, 38** -1, 28 -0, 89 _ _ _ Euromedmoins 2 _ _ _ -1, 53* -1, 24 -1, 67 Euromed 13 _ _ _ -1, 55* -1, 44 -1, 17 Euromed>=4 _ _ _ -1, 82*** -1, 63 -1, 45 Ln(PIB/hab) _ _ -1, 10 _ _ -1, 02 %Exportations Pétrolières _ _ 0, 02 Ouverture Commerciale _ _ 0, 01 NON OUI OUI 2428, 00 1258, 00 R-squared 0, 79 0, 82 0, 86 Effets Spécifiques Temps 6, 29 1, 81 13, 89 6, 26 1, 80 2, 60 Prob>F 0, 00 Euromed Y*CONST Nbre Observations 10 0, 03 0, 00

Effects on democracy using DID FH 1 2 3 4 5 6 Euromed 0, 52*** 0, 37 0, 35 _ _ _ Euromed(-2) _ _ _ 0, 44** 0, 39 0, 44 Euromed 13 _ _ _ 0, 53*** 0, 38 Euromed>=4 _ _ _ 0, 74*** 0, 57 0, 58 Ln(PIB/hab) _ _ -0, 33 _ _ -0, 35 %fuel exports _ _ -0, 01* Trade openness _ _ 0, 00 Y*CONST NO YES NO Number observations 2321, 00 1198, 00 R² 0, 78 0, 81 0, 84 Time specific effects 4, 20 4, 40 31, 00 4, 07 3, 52 30, 09 Prob>F 0, 00 0, 00 11

Impact on overall governance using within comparisons Overall governance 1 2 3 4 0, 35 0, 48** _ _ Euromed(-2) _ _ 0, 06 0, 11 Euromed 13 _ _ 0, 06 0, 38 Euromed>= 4 _ _ 0, 42 0, 52 Ln(GDP/capita) _ 1, 24 _ 1, 75 % Fuel exports _ 0, 01 _ 0, 02* Trade Openness _ 0, 01** Number of observations 120 110 140 124 R² 0, 82 0, 84 0, 81 0, 86 Time specific effects 4, 88 13, 73 40, 28 Prob>F 0, 00 Euromed 12 4, 48 0, 00

Impact on the various aspects of governance using within comparisons Bureaucratic quality Corruption Law and order 1 2 3 4 5 6 7 8 0, 34 _ -0, 05 _ 0, 80* 0, 97** _ _ Euromed(-2) _ -0, 51 _ 0, 19 _ _ 0, 38 0, 43 Euromed 13 _ -0, 43 _ 0, 15 _ _ 0, 39 0, 97* Euromed>= 4 _ -0, 32 _ 0, 28 _ _ 1, 15 1, 38* Ln(GDP/capita) _ _ _ 2, 37* _ 2, 15 % Fuel exports _ _ _ 0, 02* Trade Openness _ _ _ 0, 02*** Number of observations 120, 00 140, 00 120, 00 110, 00 140, 00 124, 00 R² 0, 68 0, 63 0, 75 0, 76 0, 81 0, 85 0, 82 0, 86 Time specific effects 6, 14 4, 50 13, 78 6, 06 4, 28 11, 12 12, 30 Prob>F 0, 00 0, 00 Euromed 13 9, 38 0, 00

Impact on governance as a whole using DID 1 2 3 4 5 6 Euromed 0, 36** 0, 42* 0, 52** _ _ _ Euromed(-2) _ _ _ 0, 28* 0, 33** 0, 36*** Euromed 13 _ _ _ 0, 29* 0, 38* 0, 48** Euromed>=4 _ _ _ 0, 58*** 0, 64** 0, 82** Ln(PIB/hab) _ _ 0, 56 _ _ 0, 53 %fuel exports _ _ 0, 00 Trade openness _ _ 0, 00 Y*CONST NO YES NO Number observations 1994, 00 1128, 00 R² 0, 74 0, 78 0, 82 Time specific effects 15, 39 5, 09 3456, 59 15, 77 5, 89 E+08 2022, 31 Prob>F 0, 00 14 0, 00 0, 00

Impact on the various components of governance Bureaucratic Quality Corruption 1 2 3 4 Law and order 6 7 8 1, 19*** 1, 11*** _ _ Euromed -0, 07 Euromed(-2) _ -0, 12 _ 0, 20 _ _ 0, 97*** 0, 77** Euromed 13 _ -0, 11 _ 0, 25 _ _ 1, 06*** 1, 10** Euromed>=4 _ -0, 07 _ 0, 38* _ _ 1, 79*** 1, 62*** Ln(PIB/hab) _ _ 1, 14** _ 1, 07** %fuel exports _ _ 0, 01 Trade openness _ _ 0, 00 Y*CONST YES YES NO YES Number observations 1994 1994 1128 R² 0, 69 0, 76 0, 71 0, 83 0, 72 0, 83 7, 44 1, 66 E+11 2, 76 3, 11 0, 00 Time specific effects Prob>F 0, 26 5 0, 00 15, 05 12761, 72 0, 00 15, 36 20400, 75 15 0, 00

Concluding remarks ¢ ¢ A strong positive effect on the quality of the judicial system; The impact on the overall governance is less clear-cut; No effect on the respect of democratic principles; Several reasons can explain this lack of results: l l l ¢ 16 EU-MED agreements are still recent the threat of sanctions by the EU is not credible the weakness of the resources dedicated to democratization Our analysis must be extended to all Trade Agreements the European Community has ratified.